Gastric cancer remains one of the most lethal malignancies worldwide, with China reporting 358,700 new cases and a mortality rate of 26.04 per 100,000 in 2022. While early-stage gastric cancer achieves 90-100% five-year survival rates after surgery, most patients present with advanced disease where conventional chemotherapy provides only 8 months median overall survival.
The advent of immune checkpoint inhibitors has transformed gastric cancer treatment, with combination immunotherapy gradually becoming standard care. However, a significant portion of patients fail to benefit from these therapies while experiencing treatment-related adverse effects, highlighting the critical need for reliable predictive biomarkers.
FDA-Approved Biomarkers Show Clinical Promise
PD-L1 Expression Emerges as Primary Predictor
PD-L1 expression, measured by combined positive score (CPS), has demonstrated consistent predictive value across multiple clinical trials. In the KEYNOTE-059 trial, patients with PD-L1-positive gastric cancer (CPS ≥ 1) receiving pembrolizumab monotherapy showed significantly higher objective response rates compared to PD-L1-negative patients.
The CheckMate-649 trial provided compelling evidence that patients with CPS ≥ 5 receiving nivolumab plus chemotherapy achieved clinically meaningful improvements in overall survival and progression-free survival compared to chemotherapy alone. Most notably, the KEYNOTE-062 trial demonstrated that pembrolizumab monotherapy in patients with CPS ≥ 10 showed significantly longer median overall survival compared to chemotherapy alone, while no significant benefit was observed at the CPS ≥ 1 threshold.
These findings establish CPS ≥ 5, and particularly CPS ≥ 10, as more definitive indicators of therapeutic benefit, though variations in patient cohort sizes and study endpoint definitions across trials necessitate further large-scale prospective studies to optimize these cutoff values.
MSI-H/dMMR Status Demonstrates Exceptional Response
Microsatellite instability-high/deficient mismatch repair (MSI-H/dMMR) tumors, representing 8-10% of gastric cancers, show remarkable sensitivity to immune checkpoint inhibitors. These tumors exhibit high tumor immunogenicity and dense immune cell infiltration due to defective DNA mismatch repair leading to abundant neoantigen generation.
In KEYNOTE-062 subgroup analysis, MSI-H gastric cancer patients receiving pembrolizumab alone or combined with chemotherapy showed significantly better objective response rates and progression-free survival compared to chemotherapy alone. A comprehensive analysis by Marabelle et al. examining 223 patients with advanced MSI-H/dMMR cancers found that gastric cancer patients achieved a 45.8% objective response rate, higher than the overall 34.3% rate, with over 75% of responders showing durable responses of ≥ 24 months.
TMB and EBV Status Show Variable Predictive Value
Tumor mutational burden (TMB) ≥ 10 mutations per megabase has received FDA approval as a tumor-agnostic biomarker for pembrolizumab. In the KEYNOTE-158 study, gastric cancer patients with TMB-H tumors achieved a 29% objective response rate compared to only 6% in non-TMB-H patients.
Epstein-Barr virus-associated gastric cancer (EBVaGC), comprising 2-10% of cases, shows variable response rates to immunotherapy. While some studies report remarkable 100% objective response rates in EBV-positive patients, others show no objective responses, highlighting the need for better patient selection criteria and understanding of resistance mechanisms.
Novel Biomarkers Expand Therapeutic Horizons
Liquid Biopsy Offers Real-Time Monitoring
Circulating tumor DNA (ctDNA) analysis has emerged as a promising tool for both predicting immunotherapy response and monitoring treatment efficacy. Studies demonstrate that patients with >25% decrease in ctDNA variant allele frequency show longer median progression-free survival (7.3 vs. 3.6 months) and higher objective response rates (53.3% vs. 13.3%).
Jin et al. found that patients with TGFBR2, RHOA, and PREX2 mutations in gastric cancer had significantly shorter progression-free survival than those without mutations. Additionally, a decrease in ctDNA concentration by the sixth week of continuous monitoring can predict immunotherapy benefit, offering potential for early treatment adjustment.
Gut Microbiota Influences Treatment Response
The gut microbiome's complex interplay with the tumor microenvironment significantly impacts immunotherapy efficacy. Research on advanced gastrointestinal cancers reveals that patients responding to PD-1/PD-L1 therapy exhibit higher relative abundance of Prevotella and Bacteroides.
Helicobacter pylori infection, common in gastric cancer patients, may hinder immunotherapy effectiveness by reducing pro-inflammatory cytokines and enhancing anti-inflammatory cytokine secretion. Retrospective analyses establish links between H. pylori infection and poorer survival outcomes in gastric cancer patients treated with immune checkpoint inhibitors.
Imaging Biomarkers Enable Non-Invasive Assessment
Radiomic features extracted from CT imaging have been validated as predictors of immunotherapy response in gastric cancer. Nuclear medicine molecular imaging techniques using 89Zr-labeled anti-PD-L1 antibodies and anti-CD8 single-domain antibodies show potential for visualizing immune-related markers in vivo.
68Ga-FAPI-04 PET/CT imaging enables non-invasive depiction of cancer-associated fibroblasts within the immunosuppressive tumor microenvironment, potentially serving as a prognostic indicator for survival and antitumor immune responses.
Overcoming Current Limitations
Despite significant progress, single biomarkers cannot completely predict immunotherapy efficacy due to temporal and spatial heterogeneity, standardization challenges, and complex immune escape mechanisms. The development of mathematical models incorporating multiple biomarkers represents a promising future direction.
Research indicates that combining TMB with T-cell-inflamed gene expression profiles achieves superior predictive performance compared to single biomarkers. In KEYNOTE clinical datasets, patients with both high TMB and high gene expression profiles achieve the highest objective response rates.
The review emphasizes that approximately 50% of MSI-H tumor patients and a considerable portion of PD-L1-positive patients still exhibit intrinsic resistance to immune checkpoint inhibitors, highlighting the urgent need for comprehensive resistance mechanism studies and alternative therapeutic strategies.
Future Directions and Clinical Implications
The integration of multi-omics approaches, including genomics, transcriptomics, and metabolomics, offers new avenues for identifying comprehensive biomarker panels. Ongoing clinical trials are investigating the predictive role of various biomarkers, including NCT05594381 examining ctDNA for evaluating PD-1 inhibitor efficacy in locally progressive gastric cancer.
The development of standardized protocols for biomarker assessment, particularly for liquid biopsy and imaging-based approaches, remains crucial for clinical implementation. Future research should focus on understanding the dynamic evolution of biomarkers during immunotherapy and their relationship with the tumor immune microenvironment.
As immunotherapy continues to evolve in gastric cancer treatment, the identification and validation of reliable predictive biomarkers will be essential for bringing precision immunotherapy to clinical practice, ultimately improving patient outcomes while minimizing unnecessary treatment-related toxicities.
